Description

伝富AI-API文档技能 - 快速查询和使用302+个AI接口，涵盖聊天、图像、视频、音频、Midjourney等多个类别

README (SKILL.md)

\r \r

伝富AI-API 调用技能\r

Name: API-Station
Author: qb-chen

\r 本技能帮助你调用伝富AI-API平台的302+个API接口。当用户需要调用AI能力时，使用此技能。\r \r

🔐 关键地址配置\r

\r

[!IMPORTANT]\r 请注意区分以下三个关键地址：\r \r | 类型 | 地址 | 说明 |\r | ----------------- | --------------------------------- | --------------------------------- |\r | API请求地址 | https://api.winfull.cloud-ip.cc | 以此开头调用所有接口 (Base URL) |\r | API官网/Token | https://api.winfull.cloud-ip.cc/ | 在此注册账户、充值、申请API Token |\r | API文档地址 | https://winfull.apifox.cn/ | 查阅最新的接口文档、参数说明 |\r \r 认证方式: 所有请求必须在Header中携带Bearer Token\r \r

Authorization: Bearer sk-xxxxxxxx\r
```\r
\r
---\r
\r
## 📚 API分类速查\r
\r
完整文档地址：https://winfull.apifox.cn/\r
\r
| 类别       | 接口数 | 说明                                |\r
| ---------- | ------ | ----------------------------------- |\r
| 图像生成   | 82     | DALL·E、Flux、豆包、Ideogram、Imagen |\r
| 视频生成   | 54     | Sora、Veo、可灵、即梦、Minimax      |\r
| 聊天       | 43     | GPT、Claude、Gemini、DeepSeek       |\r
| 任务查询   | 33     | 异步任务查询                        |\r
| 其他       | 24     | 文件上传、模型列表、代码执行等      |\r
| 函数调用   | 12     | Function Calling、Web搜索           |\r
| Replicate  | 12     | Replicate平台模型                   |\r
| 音乐生成   | 9      | Suno音乐生成                        |\r
| Midjourney | 8      | Midjourney完整功能                  |\r
| 音频       | 7      | TTS、Whisper、音频理解              |\r
| 系统API    | 7      | Token管理、用户信息                 |\r
| 可灵Kling  | 6      | 可灵专属功能                        |\r
| 文档处理   | 3      | PDF理解、文档解析                   |\r
| 嵌入       | 2      | Embeddings向量化                    |\r
\r
**总计**: 302+ 个API接口\r
\r
---\r
\r
## 🚀 核心API调用模式\r
\r
### 1. 聊天补全 (Chat Completions)\r
\r
**端点**: `POST /v1/chat/completions`\r
\r
```python\r
import requests\r
\r
response = requests.post(\r
    "https://api.winfull.cloud-ip.cc/v1/chat/completions",\r
    headers={\r
        "Authorization": "Bearer sk-xxx",\r
        "Content-Type": "application/json"\r
    },\r
    json={\r
        "model": "gpt-5.2",\r
        "messages": [\r
            {"role": "system", "content": "You are a helpful assistant."},\r
            {"role": "user", "content": "你好"}\r
        ]\r
    }\r
)\r
\r
result = response.json()\r
print(result["choices"][0]["message"]["content"])\r
```\r
\r
---\r
\r
### 2. 图像生成 (Image Generations)\r
\r
**端点**: `POST /v1/images/generations`\r
\r
```python\r
response = requests.post(\r
    "https://api.winfull.cloud-ip.cc/v1/images/generations",\r
    headers={\r
        "Authorization": "Bearer sk-xxx",\r
        "Content-Type": "application/json"\r
    },\r
    json={\r
        "model": "gpt-image-1.5-all",\r
        "prompt": "一只可爱的猫咪在草地上",\r
        "size": "1024x1536"\r
    }\r
)\r
\r
result = response.json()\r
image_url = result["data"][0]["url"]\r
```\r
\r
---\r
\r
### 3. 图片上传到图床\r
\r
> [!IMPORTANT]\r
> **图生视频必读**: 如果需要使用本地图片生成视频，必须先将图片上传到图床获取公网URL，然后使用该URL传给视频生成接口。\r
> \r
> **两种方式**:\r
> 1. 上传本地图片到图床（推荐）\r
> 2. 直接使用已有的公网图片地址\r
\r
**端点**: `POST https://imageproxy.zhongzhuan.chat/api/upload`\r
\r
```python\r
# 方式1: 上传本地图片到图床\r
with open("reference_image.jpg", "rb") as f:\r
    response = requests.post(\r
        "https://imageproxy.zhongzhuan.chat/api/upload",\r
        headers={"Authorization": "Bearer sk-xxx"},\r
        files={"file": f}\r
    )\r
\r
result = response.json()\r
image_url = result["url"]  # 获取图床URL\r
print(f"图片已上传: {image_url}")\r
\r
# 方式2: 直接使用公网图片地址\r
# image_url = "https://example.com/your-image.jpg"\r
\r
# 现在可以使用这个URL进行图生视频\r
```\r
\r
**curl示例**:\r
```bash\r
curl --location --request POST 'https://imageproxy.zhongzhuan.chat/api/upload' \\r
--header 'Authorization: Bearer \x3Ctoken>' \\r
--form 'file=@"/path/to/your/image.png"'\r
```\r
\r
---\r
\r
### 4. 视频生成 (异步任务)\r
\r
**创建视频**: `POST /v1/video/create`\r
\r
```python\r
import time\r
\r
# 如果需要使用参考图，先上传图片到图床\r
image_url = None\r
if use_reference_image:\r
    with open("reference.jpg", "rb") as f:\r
        upload_response = requests.post(\r
            "https://imageproxy.zhongzhuan.chat/api/upload",\r
            headers={"Authorization": "Bearer sk-xxx"},\r
            files={"file": f}\r
        )\r
    image_url = upload_response.json()["url"]\r
    print(f"参考图已上传: {image_url}")\r
\r
# 步骤1: 创建视频任务\r
response = requests.post(\r
    "https://api.winfull.cloud-ip.cc/v1/video/create",\r
    headers={\r
        "Authorization": "Bearer sk-xxx",\r
        "Content-Type": "application/json"\r
    },\r
    json={\r
        "model": "sora-2",\r
        "prompt": "一只小狗在草地上奔跑",\r
        "images": [image_url] if image_url else [],  # 使用图床URL\r
        "orientation": "portrait",\r
        "duration": 15\r
    }\r
)\r
\r
result = response.json()\r
task_id = result["id"]\r
```\r
\r
**查询任务**: `GET /v1/video/query?id={task_id}`\r
\r
```python\r
# 步骤2: 轮询查询任务状态\r
while True:\r
    result = requests.get(\r
        f"https://api.winfull.cloud-ip.cc/v1/video/query?id={task_id}",\r
        headers={"Authorization": "Bearer sk-xxx"}\r
    ).json()\r
    \r
    if result["status"] == "completed":\r
        video_url = result["video_url"]\r
        print(f"视频生成成功: {video_url}")\r
        break\r
    elif result["status"] == "failed":\r
        raise Exception(f"视频生成失败: {result.get('error')}")\r
    \r
    time.sleep(5)\r
```\r
\r
---\r
\r
### 5. 语音合成 (TTS)\r
\r
**端点**: `POST /v1/audio/speech`\r
\r
```python\r
response = requests.post(\r
    "https://api.winfull.cloud-ip.cc/v1/audio/speech",\r
    headers={\r
        "Authorization": "Bearer sk-xxx",\r
        "Content-Type": "application/json"\r
    },\r
    json={\r
        "model": "tts-1",\r
        "input": "你好，世界！",\r
        "voice": "alloy"\r
    }\r
)\r
\r
with open("output.mp3", "wb") as f:\r
    f.write(response.content)\r
```\r
\r
---\r
\r
### 6. 语音识别 (Whisper)\r
\r
**端点**: `POST /v1/audio/transcriptions`\r
\r
```python\r
response = requests.post(\r
    "https://api.winfull.cloud-ip.cc/v1/audio/transcriptions",\r
    headers={"Authorization": "Bearer sk-xxx"},\r
    files={"file": open("audio.mp3", "rb")},\r
    data={"model": "whisper-1"}\r
)\r
\r
result = response.json()\r
transcribed_text = result["text"]\r
```\r
\r
---\r
\r
### 7. Midjourney\r
\r
**提交Imagine任务**: `POST /mj/submit/imagine`\r
\r
```python\r
response = requests.post(\r
    "https://api.winfull.cloud-ip.cc/mj/submit/imagine",\r
    headers={\r
        "Authorization": "Bearer sk-xxx",\r
        "Content-Type": "application/json"\r
    },\r
    json={\r
        "prompt": "a beautiful sunset over the ocean --ar 16:9 --v 6"\r
    }\r
)\r
\r
result = response.json()\r
task_id = result["result"]\r
```\r
\r
---\r
\r
## 📋 异步任务通用流程\r
\r
大多数生成类API（视频、音乐、部分图像）都是异步的：\r
\r
1. 创建任务 → 获取task_id\r
2. 轮询查询状态 (pending/processing)\r
3. 状态变为completed → 获取结果URL\r
4. 状态变为failed → 处理错误\r
\r
**轮询建议**: 间隔5-10秒查询一次任务状态\r
\r
---\r
\r
## ⚠️ 重要注意事项\r
\r
1. **认证方式**: 所有请求必须在Header中携带 `Authorization: Bearer sk-xxxxxxxx`\r
2. **图生视频流程**: \r
   - 本地图片必须先上传到图床: `https://imageproxy.zhongzhuan.chat/api/upload`\r
   - 或使用已有的公网图片URL\r
   - 然后将URL传给视频生成接口的 `images` 参数\r
3. **异步轮询**: 视频/音乐生成需要轮询，间隔建议5-10秒\r
4. **Base URL**: 所有接口都以 `https://api.winfull.cloud-ip.cc` 开头\r
5. **速率限制**: 合理控制请求频率，避免触发限流\r
6. **Token计费**: 不同模型计费标准不同，详见官网\r
\r
---\r
\r
## 📖 详细文档参考\r
\r
- **在线API文档**: https://winfull.apifox.cn/\r
- **完整接口列表**: 查看 `api_list.md` 文件\r
- **官网/Token管理**: https://api.winfull.cloud-ip.cc/\r
\r
---\r
\r
## ❓ 常见问题\r
\r
### Q1: 如何获取API Token?\r
访问 https://api.winfull.cloud-ip.cc/ 注册账户，在控制台中申请API Token。\r
\r
### Q2: 如何使用本地图片生成视频?\r
必须先将本地图片上传到图床：\r
```python\r
# 1. 上传图片到图床\r
with open("image.jpg", "rb") as f:\r
    upload_resp = requests.post(\r
        "https://imageproxy.zhongzhuan.chat/api/upload",\r
        headers={"Authorization": "Bearer sk-xxx"},\r
        files={"file": f}\r
    )\r
image_url = upload_resp.json()["url"]\r
\r
# 2. 使用图片URL创建视频\r
video_resp = requests.post(\r
    "https://api.winfull.cloud-ip.cc/v1/video/create",\r
    headers={"Authorization": "Bearer sk-xxx", "Content-Type": "application/json"},\r
    json={"model": "sora-2", "prompt": "描述", "images": [image_url]}\r
)\r
```\r
\r
### Q3: 可以直接使用公网图片URL吗?\r
可以！如果图片已经在公网上，可以直接使用URL，无需上传：\r
```python\r
image_url = "https://example.com/your-image.jpg"\r
# 直接使用这个URL创建视频\r
```\r
\r
### Q4: 异步任务一直处于pending状态怎么办?\r
- 检查任务ID是否正确\r
- 等待更长时间（某些模型生成较慢）\r
- 查看账户余额是否充足\r
- 联系技术支持\r
\r
### Q5: 视频生成需要多长时间?\r
根据模型和视频长度不同，通常需要1-10分钟。建议每5-10秒轮询一次任务状态。\r

Usage Guidance

This skill is essentially a documented client for a third‑party API. Before using it: (1) verify the legitimacy and reputation of the API hosts (https://api.winfull.cloud-ip.cc and https://imageproxy.zhongzhuan.chat); (2) only upload local files you are comfortable sharing—the instructions will send images to an external image proxy to get public URLs; (3) use least-privilege or expendable API keys (do not reuse high-privilege credentials); (4) prefer HTTPS endpoints and inspect responses before sharing downstream; and (5) if you need stronger guarantees, contact the API provider or test with non-sensitive data first.

Capability Analysis

Type: OpenClaw Skill Name: ai-api-transit-station Version: 1.0.0 The skill bundle is designed to provide an AI agent with the ability to interact with the 伝富AI-API platform for various AI capabilities (chat, image, video, audio, etc.). The `SKILL.md` file serves as documentation, providing API endpoints, authentication methods, and Python/curl code examples. While the skill demonstrates the capability to read local files (e.g., `reference_image.jpg`, `audio.mp3`) and upload them to external services (`imageproxy.zhongzhuan.chat`, `api.winfull.cloud-ip.cc`), this functionality is directly tied to the stated purpose of the AI APIs (e.g., image-to-video, audio transcription). There is no evidence of intentional malicious behavior, such as data exfiltration of sensitive system files, unauthorized command execution, persistence mechanisms, or deceptive prompt injection attempts against the agent. The instructions are clear and align with the skill's stated purpose.

Capability Assessment

✓ Purpose & Capability

Name/description claim a catalogue of AI endpoints (chat, image, video, TTS, etc.) and the SKILL.md provides concrete endpoints and examples covering those categories. The skill does not request unrelated credentials, binaries, or installs, which matches the stated purpose of being an API-invocation helper.

ℹ Instruction Scope

Runtime instructions are focused on calling the documented API endpoints. Sample code shows reading local image files and uploading them to a remote image proxy (imageproxy.zhongzhuan.chat) when generating video — this is coherent with the feature (image→video) but means the agent (or user code) will send local files to an external host. No instructions ask the agent to read unrelated system files or secrets.

✓ Install Mechanism

There is no install spec and no code files; the skill is instruction-only, so nothing is written to disk and there is no package download risk.

✓ Credentials

The skill declares no required environment variables or credentials. The SKILL.md uses an Authorization: Bearer sk-xxx pattern in examples (the user must supply an API token to call the service), which is proportional and expected for an API client. The skill does not request unrelated secrets or cross-service credentials.

✓ Persistence & Privilege

The skill is not always-enabled and does not request elevated/persistent privileges. It does not attempt to modify other skills or system-wide configuration.

Version History

v1.0.0

A brand new AI service platform, unlocking the door to the future, AI aggregation management. The most stable official enterprise-level channels across the entire network, with prices as low as 0.95% of the official rate. Register and get $0.2 instantly, with detailed records available; every expenditure is open and transparent.

Metadata

Slug ai-api-transit-station

Version 1.0.0

License —

All-time Installs 0

Active Installs 0

Total Versions 1

Frequently Asked Questions

What is API-Station?

伝富AI-API文档技能 - 快速查询和使用302+个AI接口，涵盖聊天、图像、视频、音频、Midjourney等多个类别. It is an AI Agent Skill for Claude Code / OpenClaw, with 468 downloads so far.

How do I install API-Station?

Run "/install ai-api-transit-station" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is API-Station free?

Yes, API-Station is completely free (open-source). You can download, install and use it at no cost.

Which platforms does API-Station support?

API-Station is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created API-Station?

It is built and maintained by QB-Chen (@qb-chen); the current version is v1.0.0.

More Skills

API-Station